Towards Automatic Document Classification Using a Fuzzy Logic Approach

Nowadays organizations need technology to improve processes, reduce operations time, minimize potential human errors, and maximize the benefits. In this sense, document classification became a priority. Artificial Intelligence (AI) is a field that contributes to the automation of processes. This research aims to find a feasible solution to automating classifying documents. Several studies aim to classify data, but most of them are applied to numerical variables. Nevertheless, there is a methodology that could express in the results a gradient of similarities with certain topics, deals well with uncertainty, and is believed to be a potential solution to this problem: Fuzzy Logic. Although it seems to be a possible solution to the problem in question, the results attained and shown here were not as successful as expected. This research compares the results of a model created using Fuzzy Logic for topic classification with other models identified from the state of the art, which have been used for the same purpose, but in any case, the author of this paper suggests not disregarding this technique to accomplish a new methodology in automatic document classification and propose some modifications that can make a model created using this technique attain enhanced performance.